How Did This Get Funded?! Automatically Identifying Quirky Scientific Achievements

Abstract

Humor is an important social phenomenon, serving complex social and psychological functions. However, despite being studied for millennia humor is computationally not well understood, often considered an AI-complete problem.In this work, we introduce a novel setting in humor mining: automatically detecting funny and unusual scientific papers. We are inspired by the Ig Nobel prize, a satirical prize awarded annually to celebrate funny scientific achievements (example past winner: "Are cows more likely to lie down the longer they stand?"). This challenging task has unique characteristics that make it particularly suitable for automatic learning.We construct a dataset containing thousands of funny papers and use it to learn classifiers, combining findings from psychology and linguistics with recent advances in NLP. We use our models to identify potentially funny papers in a large dataset of over 630,000 articles. The results demonstrate the potential of our methods, and more broadly the utility of integrating state-of-the-art NLP methods with insights from more traditional disciplines.

Simple and Effective Multi-Token Completion from Masked Language Models

“…Transformer-based LMs and MLMs (Peters et al, 2018;Devlin et al, 2018) have revolutionized NLP in the past couple of years. While most of the impact has been achieved using these pretrained models as a source of meaningful contextual embeddings, recent works are using these models for the task they were pretrained for: Masked Language Modeling (Petroni et al, 2019;Kushilevitz et al, 2020;Lazar et al, 2021;Shani et al, 2021;Jiang et al, 2020).…”

Transformers Go for the LOLs: Generating (Humourous) Titles from Scientific Abstracts End-to-End

“…The majority were labeled as non-funny and annotators exhibited low agreements. Shani et al (2021) classify scientific titles as funny or not using humor-theory inspired features and scientific language models such as SciBERT (Beltagy et al, 2019) building on a dataset of Ig Nobel winners and humorous papers discussed in online forums.…”

“…To generate humorous titles, we first need a dataset of humor annotated titles in our domain (NLP and ML papers). We cannot resort to the data of Shani et al (2021); Heard et al (2022) as those leverage papers from other scientific fields. As a consequence, we build our own dataset.…”